Journul of Chromatography, Biomedical
519
(I 992) 73-83
Applications
Elsevier Science Publishers
CHROMBIO.
B.V., Amsterdam
6414
Monitoring of an experimental red blood cell pathology with gravitational field-flow fractionation A. Merino-Dugay,
Ph. J. P. Cardot,
M. Czok and M. Guernet
Lohoratoire de Chimie Analytique et d’Electrochimie Organique, Ckment, F-92296 Chcitenay-Malabry Cedex (France)
Centre d’Etudes Pharmaceutiques,
Universite Paris-&d,
5 Rue J. B.
J. P. Andreux Laborutoire d’Hmatologie, Malabry Cedes (France)
(First received
Centre d’Etudes Phurmaceutiyues,
December
26th, 1991; revised manuscript
Universit6 Paris-&d,
5 Rue J. B. Ckment,
F-92296 Chdtenoy-
received April 13th, 1992)
ABSTRACT Gravitational
field-flow fractionation
blood cells (RBC).
The variation
induced by phenylhydrazine
is a simple method
in the composition
was studied.
Blood samples taken at different
shape of the elution profiles. Microscopic channel
make a description
reticulocytes, 96 to
observations
of the RBC population
and the blood cells containing
of original possible.
studied.
after the second day of the anaemia
Granulometric
studies were performed
fractionation.
relaxation
process of these cells was then studied to approximate
different
stop-flow
microscopic
observation
subpopulations regression
size were selectively
times enabled
the study
and size analysis,
eluted,
of the anaemia
Correspondence
process
permit the description The correlation
Analytique
Laboratoire Centre d’Etudes
Cedex, France.
of modifications
monitoring
Chimie
between
for example the red reversible
density
anaemia
in the retention
and
at the outlet of the fractionation
were the normal
RBC, the newly produced
A decrease
of the normal
by the anaemia
RBC from
and increases
in
on the third day, were also
each day as well as on some fractions
of known
on the elution
observed
blood fractionation (FFF) years [1,2].
to: Dr. Ph. J. P. Cardot,
collected
is stimulated
their density properties.
INTRODUCTION
in field-flow
from the rabbit
as were cells of equivalent
opens a new field in the analytical
Studies
showed differences
precipitates).
in bone marrow
of some cell subpopulations
of the relaxation
at each stage of the anaemia.
haemoglobin
particles,
an experimental
1 to 16%). RBC with Heinz bodies, which appear
on the RBC sampled
procedures
during
stages of the aenemia
The three species observed
of reticulocytes
field-flow
but of the same average
Reinjection
(from
of micrometre-size
in a rabbit
RBC samples and fractions
Heinz bodies (intracellular
1% was observed over five days. The production
percentage
suitable for the separation
of the RBC population
size distribution
eluted by
were also performed.
The
It was observed
that RBC of different density
but of different
size. Injection
profiles.
The results,
in the RBC composition the fractionation
compared
of the cells at with systematic
as well as the purification
profiles
and the progress
of this type of pathology.
With some restrictions, fundamental experience with particles other biological
size, density and
eluted if they differ [3-51.
Phar-
tions has also been demonstrated shown that size and density and possibly rights reserved
of
or the
74
shape, were involved. It is then possible to monitor modifications of cell populations by FFF techniques and to correlate the characteristics of the cells with their elution profiles. A well established, reversible model of red cell change is provided by haemolytic anaemia induced in rabbits by phenylhydrazine [7]. This process, by simultaneous steps of haemolysis of the RBC in the peripheral circulation and regeneration in the haematopoi’etic bone marrow, will modify the RBC proportions of four characteristic subpopulations, which are easy to characterize by microscopic observation. The four populations are: normal RBC, reticulocytes and two types of cell that include in their cytoplasma Heinz bodies (haemoglobin precipitates) at levels below or above ten inclusions per cell. Investigation of this anaemia with the FFF can open opportunities in purification, and allow an type of of the of RBC The principles FFF elution have been described [335,8,9]. date, three models are (normal, steric inertial, also hyperlayer). It also been that the of RBC gravitational FFF to be process that both the and the ertial models In such process, it been demonstrated the separation on both density and size. Thereit is to separate of the density and different size well as with the average size of different ty* We investigated the FFF behaviour living rabbit during an day period. anaemia was the first days by In the circulation, induces an denaturation haemoglobin. This becomes visible the presence Heinz bodies the RBC. cells do transport any tional haemoglobin the circulation, they are to be trapped and in the The strong of functional moglobin levels the production new
A. Merino-Dugay
rt al. / J. C’hromarogr.
579 (1992)
73-83
RBC the bone This stimulation evidenced an increase the number very young called the Under phellylhydrazine the turnover RBC is by a destruction and of cells, modifies the composition during after the leading a population an average lower than the treatment. elution experimicroscopic observations Coulter counter measurements were to monitor RBC cell during the anaemia. EXPERIMENTAL
Fieldyjlow jkactionation The separation channel was built similarly to the ones already described by Giddings and Myers [l I]. Two mirror-quality glass plates, which enclosed a Mylar band, were compressed by two screwed plexiglass plates. The Mylar band was cut to produce a rectangular channel. To meet the special conditions required for living biological material, the glass plates (channel walls) were coated with biocompatible silicone (Silbione, Rhone Poulenc, Paris, France). The void volume of the channel, including the tubing connections, is 3.16 ml, and the dimensions of the channel were 85.5 x 2 x 0.0175 cm. The carrier mobile phase and sample dilution solvent was an isotonic phosphate buffer (pH 7.4). Most of the experiments described in this report were performed with the same injection protocol. In the abscence of flow, 100 or 50 ,nl of the suspension containing from cu. 200 000 to ca. 400 000 RBC were injected through a septum device in the channel. The particles settled during a given time (usually 4 min) in the thickness of the channel, then the flow was established using a Waters 6OOOA pump. Detection was performed with a photometric detector set at 313 nm (Shimadzu SP6A). The fractogram parameters were determined by classical methods [ 1,2,12]. The peak profile accuracy and the precision in work with biological materials were established with human RBC of known characteristics. The results
A. Merino-&gay
TABLE PEAK
et al. 1 1. Chromatogr.
579 (1992)
75
73-83
I CHARACTERISTICS:
Channel
thickness,
PRECISION
0.175 mm; injection
Reproducibility
MEASUREMENTS
with relaxation;
in 50 ~1; precision,
200 000 particles
over 24 h”
f
20.
Day-to-day -..-
Retention
factor
Peak width
Asymmetry
factor
Latex beads
0.15 f 0.002
5.2 f 0.10
1.6 f 0.08
Red blood cells
0.28 f 0.03
2.4 zt 0.05
0.93 f 0.020
Peak height I5 f 0.8
a Latex beads flow-rate,
0.69 cm/s; RBC flow-rate,
0.43 cm/s; rz = 4.
b Latex beads Bow-rate,
0.74 cm/s; RBC flow-rate,
0.43 cm/s; rr = 14 (two measurements
of this validation process are shown in Table I, in comparison with latex beads of equivalent particle volume. Red blood cells Haemolytic anaemia was induced in a New Zealand white rabbit (2.24 kg) by daily subcutaneous injections ofphenylhydrazine (8.75 mg/kg) for four days. The sample of day 1 was taken just before the first drug injection, and other samples were taken on days 3, 4, 5 and 8. Each day, the blood was drawn into tubes containing EDTA as an anticoagulant. The cell volume distribution was analysed, using the resistivity detector of a Model ZM multi-channel Coulter counter set for 64-channel analysis (Coultronics, Margency, France). All the size histograms included in this report show the number of cells expressed as a percentage of the total analysed population. The average concentration of cells in the blood stream was evaluated each day by the measure of the haematocrit ratio (volume of the cell population to the sample volume). This procedure allowed the calculation of appropriate dilutions for Coulter counting. With the use of direct optical microscopy, a cell count was made on blood films stained with May-Grunewald-Giemsa [ 131 or Crystal Violet [13,14] (cells with Heinz bodies) or Brilliant Cresyl Blue (13,14] (reticulocytes). RESULTS
AND
Reversible
DISCUSSION
anaemia
induces modifications
in
Retention
reproducibility* _.________ factor
0.19 + 0.007 0.28 i 0.012
per day for seven days).
the composition of the RBC population, which can be described by classical haematological methods, such as morphological description under microscopical observation and size analysis by Coulter counting. This report shows that complementary information is obtained when the anaemia is monitored with these classical methods and with FFF. The separation power of FFF means that the purification can be optimized by fraction collection and reinjection. Modifications of retention and shape characteristics of e&ion projiles During haemolysis, by the action of phenylhydrazine and concurrent regeneration, fractograms of RBC were obtained as shown in Fig. 1. The first two peaks are observed, in all the fractograms: they can be considered as “system peaks”. They generally correspond to species not affected by the external field and their origin has been explained elsewhere [2,15]. The third peak, which can appear bimodal, corresponds to the REX. The composition variation of the RBC for eight days was described for the four most important subpopulations, which are known to be strongly modified in number by the anaemia: the normal RBC, the reticulocytes and the two populations showing either more or less than ten Heinz bodies per cell. The composition modifications of these subpopulations during the regenerative anaemia are described in Fig. 2. On day 1, the RBC were eluted as a single peak (Fig. l), and the global analysis of the blood sam-
ple showed the presence of 96% of normal RBC (Fig. 2), the remaining 4% being reticulocytes and old RBC with a single Heinz body. This symmetrical profile is characteristic of a normal RBC population and matches those published previously for both human [1,2] and animal [2,16] WY?,
SIZE
DISTRIBUTION
ON
5300
CELLS
30
i
0
0
1. Fractograms
of rabbit
phenylhydrazine-induced
RBC before,
anaemia.
DISTRIBUTION
ON
DAY 4,
SiZE
DISTRIBUTION
ON
7410
CELLS
5965
CELLS
during
and after
A lOO-~1 sample of RBC sus-
pension
of 4000 cells per ~1 injected with a 4-min stop-flow phase, isotonic phosphate
is 6.5
SIZE
60 min
Carrier
cm/s: detection
3,
60 min
30 min
Fig.
30 min
DAY
wavelength,
10e3 absorbance
time.
buffer (pH 7.4); flow-rate,
0.21
s
3 13 nm. Abs shown on fractograms
1
DAY 5,
SIZE
DISTRIBUTION
ON
5700
CELLS
30
units.
I CELLS
IN
DAY 8,
1
0
2
3
4
5
6
7
SIZE
DISTRIBUTION
ON
5750
CELLS
8 DAYS
Fig. 2. Composition
modifications
of RBC population
during
the experimental haemolytic anaemia. RBC samples were analyscd before FFF elution. Samples were stained with May Grun-
PARTICLE
Fig.
3.
Sire distribution
VOLUME
of rabbit RBC
during
haemoly-
ZM set for a 64-channel
sampled
analysis.
wald Ciiemsa. Crystal Violet and Brilliant Cresyl Blue. Subpopulation characterization was performed by statistically signifi-
sis. Coulter
counter
cant
Calibration
of the system with S.Gpm-diameter
direct
(X 100).
counting
with
optical
microscopy
magnification
Model
(pm3)
latex particles.
A. Merino-Dugay
et al. / J. Chromatogr.
579 (1992)
73-83
cells. In these elution conditions, the morphological observation of peak fractions (beginning, middle, tail) by microscopical observation did not show any enrichment in either subpopulation [17]. Under the elution conditions described in the legend of Fig. 1, a retention factor of 0.21 was obtained for the RBC peak eluted on day 1. The retention factor (R) is defined as the void volume of the separation system divided by the retention volume. On the first three days, because the phenylhydrazine-induced anaemia has been acting for 48 h, the haematocrit measurement showed a decrease of 5% per day of the RBC population. On day 3, the normal RBC population had dropped from 96% on the first day to only 6%, and 89% of the cells contained denatured haemoglobin. In 37% of the cells between one and ten Heinz bodies had been formed, and 52% contained more than ten. The elution of RBC sampled that day (day 3) produced a fractogram with two badly resolved peaks, one with a retention factor of 0.24 and the other with R = 0.17 (Fig. 1). A morphological analysis of the cell composition of the less retained peak (R = 0.24) showed essentially cells with one Heinz body (60% of the cells counted in the fraction) and a non-negligible proportion of reticulocytes. The second peak (R = 0.17) was noticeably broader and contained all the other species. Two thirds of this fraction were made up of species with more than one Heinz body. On day 4, the total blood composition was similar to that of day 3, but the majority of the Heinz cells showed more than eight granulations. It was also observed, on the fractogram of RBC sampled that day, that the height and shape of the first RBC peak had been modified. On day 5, 24 h after the phenylhydrazine injections had been stopped, there was a significant increase in the reticulocyte concentration. Fraction analysis of the bimodal peak eluted that day indicated that these reticulocytes were found primarily in the first peak. In the second peak, cells with Heinz bodies were observed and represent cu. 80% of the cell population found in that peak.
17
On day 8, the composition of the blood sample before FFF elution contained 30% of normal RBC and 20% of reticulocytes, most of the other cells containing less than ten Heinz bodies. The fractogram profile was analogous to that of day 1. It was observed that the retention factor (R = 0.31) of this rabbit RBC population was higher than that measured on day 1. Because modifications in the RBC composition caused the appearance of new peaks, retention modifications of RBC peaks, and modifications in the peak shape, a complementary analysis of the cells before and after elution was performed. In haematology and in particle analysis, size estimations are widely used, but density evaluations may also be needed [I 81. Selective size elation of RBC subpopulations Size analysis of the RBC populations was performed before elutions and in some cases after collection of fractions (day 4, day 5). Fig. 3 shows the histograms of the size distributions of the RBC sampled on days 1,3,4, 5 and 8; the volume distribution is given in percentage of the cell number. One can observe day-to-day variations in the granulometric distribution of the RBC. It can be observed that the RBC of day 8 have an average size (120 pm3) larger than those of day 1 (71 pm3). Fig. 3 shows that on day 8 the volume distribution is shifted to larger volumes. These results have to be combined with the fact that the retention factor is higher on day 8 than on day 1 (R = 0.31 and 0.21, respectively). If factors such as the density and shape or rigidity of the cells are ignored, we can conclude that in this case the cells of larger average volume (day 8) are eluted faster than the smaller ones (day 1). On day 4, a bimodal peak profile is observed and two fractions are collected, the first corresponding to the cells eluted at a retention factor of 0.24 (start 20 min, end 30 min) and the second to the cells eluted from 30 to 60 min (retention factor 0.17). The size histogram of the original sample injected is shown in Fig. 3 (day 4) and gives an average volume for the cells of 71 pm3. Note that more than 50% of the cells were in the range 60-90 pm3. Fig. 4A shows the size histo-
A. Merino-Dugay
rt al. 1 J. Chromarogr.
A DAY 4,
SIZE
579 (1992)
73-83
B
DISTRIBUTION
ON 5965
CELLS
DAY 5,
SIZE
DISTRIBUTION
40
67
ON 5700
CELLS
30 1
13
27
40
54
67
80
94
107
120
134
147
161
174
187
214
227
‘:
13
27
54
80
94
107
120 i34
147
161
174
I.97
214
227
ii u
2 ?
DAY 4,
SIZE
DISTRIBUTION
OF ELUTED
d
CELLS
6 b
DAY 5,
SIZE
DISTRIBUTION
OF ELUTED
CELLS
n Fraction15-30min, Sizeccmnton4000 cells % 0 30 z-
20
0
Fraction35-60min, Sizecounton 7478cells
20
10
0
0 13
27
40
54
67
80
PARTICLE Fig. 4. Size distribution
94
107
120
134
VOLUME
147 161
latex particles.
tions of the two purified
fractions
214
227
13
27
40
54
67
80
PARTICLE
RBC sampled Coulter
The upper graphs
collected
187
(pm3)
of RBC. (A) Rabbit
on day 5, before (upper) and after (lower) elution. with 5.6-pm-diameter
174
94
107
120 134
147
161
Model ZM set for a 64-channel
show the sample before injection,
analysis.
227
(I_‘m3)
VOLUME
before (upper) and after (lower) elution on day 4. (B) Rabbit counter
174 187214
Calibration
and the lower graphs
RBC sampled of the system
show the size distribu-
after FFF elution.
gram of the two FFF-eluted RBC subpopulations, compared with the original one. On day 4, size evaluation of both populations separated by FFF give an average volume of 71 pm3 for the population collected first (retention factor 0.24) and an average volume of 74 pm3 for the fraction eluting later: in both cases, more than 50% of the cells were in the range 60-90 pm3. This average volume difference is hard to detect by visual inspection of the figure and it may not be significant. Because of the presence again of a bimodal peak on day 5 two fractions were collected, the first including the cells eluted between 15 min and 30 min and the second between 3.5 min and 60 min. Size analysis of both fractions was performed and the size distribution histograms are
shown in Fig. 4B. It was observed that the two fractions show differences in their size distributions. The cells eluted first (15530 min) have an average volume of 120 pm3 in a broad distribution, with more than 50% of the cells in the range 100-160 pm3. The cells eluted after 35 min were smaller, with an average volume of 67 pm3. These results demonstrate that the selectivity in the retention of RBC can be due to differences in size, as also indicated by the comparison of retentions obtained on day 1 and on day 8. The elution process appeared to follow a size-dependent elution order, often described as “steric” [l], in which the bigger particles were eluted before the smaller ones. The size measurements performed on day 4, however, do not show similar results, which in-
A. Merino-Dugay
et al. / J. Chromarogr.
579 (1992)
79
73-83
dicates that the size cannot be the only parameter involved. As described previously [2,6,19], density can be taken into account, and was to be evaluated for a more complete comprehension of the elution mechanism of the RBC. Classical methods of density determination are related to density centrifugations [ZO]. Since these experiments take a relatively long time, subjecting the cells to stress and possibly jeopardizing recovery, we have tried to compare the density of the cells by FFF using the concept of “relaxation time” [21]. Relaxation and particle characterization One critical step in achieving a separation by FFF is the relaxation process [21,22]. To permit particles to reach their equilibrium position in the separation channel, it is recommended that the samples are injected into the mobile phase at stopped flow and that the flow is resumed after a period of a few minutes. The time needed to allow complete relaxation depends on the channel thickness, the particle size and density, and the carrier phase density. It is possible to determine the relaxation time of a given particle by a series of experiments performed at different stop-flow times. As the relaxation time is used to describe the transversal equilibrium state reached by the cells in the channel, a stop-flow period longer than the relaxation time will not produce any variations in the elution characteristics. On the other hand, as long as the elution profile changes with stop-flow times, one can consider that this equilibrium state has not yet been reached. The experimental relaxation time can be defined as the shortest stopflow time where no more modifications of the peak characteristics occur [23]. Three parameters have been studied on human and rabbit RBC fractograms: the retention factor, the peak width and the asymmetry factor. Results obtained with human and rabbit RBC are shown in Fig. 5 and 6. To determine the equilibrium position, a weighted first-order polynomial least-squares fitting procedure has been used (R = ,f( 1/t))_ Because of the large number of experiments (eighteen eiutions in the case of human RBC) and to empha-
size the importance of high stop-flow time experiments, the weighting procedure was time-dependent. The retention factor, the asymmetry ratio and the peak width can be extrapolated to obtain the limiting retention factor, the limiting asymmetry ratio and the limiting peak width to an infinite stop-flow time. These values are plotted in Figs. 5 and 6. Even if altered by possible systematic errors (related to equilibrium and sedilnentation) these values will be considered in this
Retention 0.6
Factor
7
Asymmetry Factor (B/A) 1.4.
1.2 1.0-
.
0.8 -
.
.
--
.
.
I
.
Limit Value=
1.15
0.6 4 0.4 0.2
. .
I
1
0.04 0
I
1
2
3
I
I
I
4
5
6
Stop Flow Time (minf
Fig. 5. Fractogram elution characteristics of normal human RBC obtained at different stop-flow times. The retention factor is calculated
as the ratio of the elution time of non-retained
cies (proteins}
to the elution
time of the summit
spe-
of the peak.
Asymmetry factor ratio (B/A) is measured at 20% of peak height. Peak width is given in minutes and measured at 50% of the peak height. human buffer. nm3.
Injection
conditions:
50 ii1 of a suspension
of
RBC containing 4500 cells per 8~1diluted in an isotonic Flow-rate 0.48 cm/s. Average volume of RBC used, 90
Retention
0.0
Factor
“S’f’I”I”1”I
0
4
1
6
8
10
Peak Width
1
6: 4-
ation time values of 1.05, 2.1, 2.2 and 2.55 min, respectively, and the average relaxation value was 1.98 i 0.6 min: It is then possible to compare these experimental values with the ones calculated using the analytical expression of the sedimentation time, described in FFF by Giddings el al. [Zl]. For human RBC, the values given in the literature [14,15,I~]areadensity(~~)of 1.086
519
(I 992) 73-83
Applications
Elsevier Science Publishers
CHROMBIO.
B.V., Amsterdam
6414
Monitoring of an experimental red blood cell pathology with gravitational field-flow fractionation A. Merino-Dugay,
Ph. J. P. Cardot,
M. Czok and M. Guernet
Lohoratoire de Chimie Analytique et d’Electrochimie Organique, Ckment, F-92296 Chcitenay-Malabry Cedex (France)
Centre d’Etudes Pharmaceutiques,
Universite Paris-&d,
5 Rue J. B.
J. P. Andreux Laborutoire d’Hmatologie, Malabry Cedes (France)
(First received
Centre d’Etudes Phurmaceutiyues,
December
26th, 1991; revised manuscript
Universit6 Paris-&d,
5 Rue J. B. Ckment,
F-92296 Chdtenoy-
received April 13th, 1992)
ABSTRACT Gravitational
field-flow fractionation
blood cells (RBC).
The variation
induced by phenylhydrazine
is a simple method
in the composition
was studied.
Blood samples taken at different
shape of the elution profiles. Microscopic channel
make a description
reticulocytes, 96 to
observations
of the RBC population
and the blood cells containing
of original possible.
studied.
after the second day of the anaemia
Granulometric
studies were performed
fractionation.
relaxation
process of these cells was then studied to approximate
different
stop-flow
microscopic
observation
subpopulations regression
size were selectively
times enabled
the study
and size analysis,
eluted,
of the anaemia
Correspondence
process
permit the description The correlation
Analytique
Laboratoire Centre d’Etudes
Cedex, France.
of modifications
monitoring
Chimie
between
for example the red reversible
density
anaemia
in the retention
and
at the outlet of the fractionation
were the normal
RBC, the newly produced
A decrease
of the normal
by the anaemia
RBC from
and increases
in
on the third day, were also
each day as well as on some fractions
of known
on the elution
observed
blood fractionation (FFF) years [1,2].
to: Dr. Ph. J. P. Cardot,
collected
is stimulated
their density properties.
INTRODUCTION
in field-flow
from the rabbit
as were cells of equivalent
opens a new field in the analytical
Studies
showed differences
precipitates).
in bone marrow
of some cell subpopulations
of the relaxation
at each stage of the anaemia.
haemoglobin
particles,
an experimental
1 to 16%). RBC with Heinz bodies, which appear
on the RBC sampled
procedures
during
stages of the aenemia
The three species observed
of reticulocytes
field-flow
but of the same average
Reinjection
(from
of micrometre-size
in a rabbit
RBC samples and fractions
Heinz bodies (intracellular
1% was observed over five days. The production
percentage
suitable for the separation
of the RBC population
size distribution
eluted by
were also performed.
The
It was observed
that RBC of different density
but of different
size. Injection
profiles.
The results,
in the RBC composition the fractionation
compared
of the cells at with systematic
as well as the purification
profiles
and the progress
of this type of pathology.
With some restrictions, fundamental experience with particles other biological
size, density and
eluted if they differ [3-51.
Phar-
tions has also been demonstrated shown that size and density and possibly rights reserved
of
or the
74
shape, were involved. It is then possible to monitor modifications of cell populations by FFF techniques and to correlate the characteristics of the cells with their elution profiles. A well established, reversible model of red cell change is provided by haemolytic anaemia induced in rabbits by phenylhydrazine [7]. This process, by simultaneous steps of haemolysis of the RBC in the peripheral circulation and regeneration in the haematopoi’etic bone marrow, will modify the RBC proportions of four characteristic subpopulations, which are easy to characterize by microscopic observation. The four populations are: normal RBC, reticulocytes and two types of cell that include in their cytoplasma Heinz bodies (haemoglobin precipitates) at levels below or above ten inclusions per cell. Investigation of this anaemia with the FFF can open opportunities in purification, and allow an type of of the of RBC The principles FFF elution have been described [335,8,9]. date, three models are (normal, steric inertial, also hyperlayer). It also been that the of RBC gravitational FFF to be process that both the and the ertial models In such process, it been demonstrated the separation on both density and size. Thereit is to separate of the density and different size well as with the average size of different ty* We investigated the FFF behaviour living rabbit during an day period. anaemia was the first days by In the circulation, induces an denaturation haemoglobin. This becomes visible the presence Heinz bodies the RBC. cells do transport any tional haemoglobin the circulation, they are to be trapped and in the The strong of functional moglobin levels the production new
A. Merino-Dugay
rt al. / J. C’hromarogr.
579 (1992)
73-83
RBC the bone This stimulation evidenced an increase the number very young called the Under phellylhydrazine the turnover RBC is by a destruction and of cells, modifies the composition during after the leading a population an average lower than the treatment. elution experimicroscopic observations Coulter counter measurements were to monitor RBC cell during the anaemia. EXPERIMENTAL
Fieldyjlow jkactionation The separation channel was built similarly to the ones already described by Giddings and Myers [l I]. Two mirror-quality glass plates, which enclosed a Mylar band, were compressed by two screwed plexiglass plates. The Mylar band was cut to produce a rectangular channel. To meet the special conditions required for living biological material, the glass plates (channel walls) were coated with biocompatible silicone (Silbione, Rhone Poulenc, Paris, France). The void volume of the channel, including the tubing connections, is 3.16 ml, and the dimensions of the channel were 85.5 x 2 x 0.0175 cm. The carrier mobile phase and sample dilution solvent was an isotonic phosphate buffer (pH 7.4). Most of the experiments described in this report were performed with the same injection protocol. In the abscence of flow, 100 or 50 ,nl of the suspension containing from cu. 200 000 to ca. 400 000 RBC were injected through a septum device in the channel. The particles settled during a given time (usually 4 min) in the thickness of the channel, then the flow was established using a Waters 6OOOA pump. Detection was performed with a photometric detector set at 313 nm (Shimadzu SP6A). The fractogram parameters were determined by classical methods [ 1,2,12]. The peak profile accuracy and the precision in work with biological materials were established with human RBC of known characteristics. The results
A. Merino-&gay
TABLE PEAK
et al. 1 1. Chromatogr.
579 (1992)
75
73-83
I CHARACTERISTICS:
Channel
thickness,
PRECISION
0.175 mm; injection
Reproducibility
MEASUREMENTS
with relaxation;
in 50 ~1; precision,
200 000 particles
over 24 h”
f
20.
Day-to-day -..-
Retention
factor
Peak width
Asymmetry
factor
Latex beads
0.15 f 0.002
5.2 f 0.10
1.6 f 0.08
Red blood cells
0.28 f 0.03
2.4 zt 0.05
0.93 f 0.020
Peak height I5 f 0.8
a Latex beads flow-rate,
0.69 cm/s; RBC flow-rate,
0.43 cm/s; rz = 4.
b Latex beads Bow-rate,
0.74 cm/s; RBC flow-rate,
0.43 cm/s; rr = 14 (two measurements
of this validation process are shown in Table I, in comparison with latex beads of equivalent particle volume. Red blood cells Haemolytic anaemia was induced in a New Zealand white rabbit (2.24 kg) by daily subcutaneous injections ofphenylhydrazine (8.75 mg/kg) for four days. The sample of day 1 was taken just before the first drug injection, and other samples were taken on days 3, 4, 5 and 8. Each day, the blood was drawn into tubes containing EDTA as an anticoagulant. The cell volume distribution was analysed, using the resistivity detector of a Model ZM multi-channel Coulter counter set for 64-channel analysis (Coultronics, Margency, France). All the size histograms included in this report show the number of cells expressed as a percentage of the total analysed population. The average concentration of cells in the blood stream was evaluated each day by the measure of the haematocrit ratio (volume of the cell population to the sample volume). This procedure allowed the calculation of appropriate dilutions for Coulter counting. With the use of direct optical microscopy, a cell count was made on blood films stained with May-Grunewald-Giemsa [ 131 or Crystal Violet [13,14] (cells with Heinz bodies) or Brilliant Cresyl Blue (13,14] (reticulocytes). RESULTS
AND
Reversible
DISCUSSION
anaemia
induces modifications
in
Retention
reproducibility* _.________ factor
0.19 + 0.007 0.28 i 0.012
per day for seven days).
the composition of the RBC population, which can be described by classical haematological methods, such as morphological description under microscopical observation and size analysis by Coulter counting. This report shows that complementary information is obtained when the anaemia is monitored with these classical methods and with FFF. The separation power of FFF means that the purification can be optimized by fraction collection and reinjection. Modifications of retention and shape characteristics of e&ion projiles During haemolysis, by the action of phenylhydrazine and concurrent regeneration, fractograms of RBC were obtained as shown in Fig. 1. The first two peaks are observed, in all the fractograms: they can be considered as “system peaks”. They generally correspond to species not affected by the external field and their origin has been explained elsewhere [2,15]. The third peak, which can appear bimodal, corresponds to the REX. The composition variation of the RBC for eight days was described for the four most important subpopulations, which are known to be strongly modified in number by the anaemia: the normal RBC, the reticulocytes and the two populations showing either more or less than ten Heinz bodies per cell. The composition modifications of these subpopulations during the regenerative anaemia are described in Fig. 2. On day 1, the RBC were eluted as a single peak (Fig. l), and the global analysis of the blood sam-
ple showed the presence of 96% of normal RBC (Fig. 2), the remaining 4% being reticulocytes and old RBC with a single Heinz body. This symmetrical profile is characteristic of a normal RBC population and matches those published previously for both human [1,2] and animal [2,16] WY?,
SIZE
DISTRIBUTION
ON
5300
CELLS
30
i
0
0
1. Fractograms
of rabbit
phenylhydrazine-induced
RBC before,
anaemia.
DISTRIBUTION
ON
DAY 4,
SiZE
DISTRIBUTION
ON
7410
CELLS
5965
CELLS
during
and after
A lOO-~1 sample of RBC sus-
pension
of 4000 cells per ~1 injected with a 4-min stop-flow phase, isotonic phosphate
is 6.5
SIZE
60 min
Carrier
cm/s: detection
3,
60 min
30 min
Fig.
30 min
DAY
wavelength,
10e3 absorbance
time.
buffer (pH 7.4); flow-rate,
0.21
s
3 13 nm. Abs shown on fractograms
1
DAY 5,
SIZE
DISTRIBUTION
ON
5700
CELLS
30
units.
I CELLS
IN
DAY 8,
1
0
2
3
4
5
6
7
SIZE
DISTRIBUTION
ON
5750
CELLS
8 DAYS
Fig. 2. Composition
modifications
of RBC population
during
the experimental haemolytic anaemia. RBC samples were analyscd before FFF elution. Samples were stained with May Grun-
PARTICLE
Fig.
3.
Sire distribution
VOLUME
of rabbit RBC
during
haemoly-
ZM set for a 64-channel
sampled
analysis.
wald Ciiemsa. Crystal Violet and Brilliant Cresyl Blue. Subpopulation characterization was performed by statistically signifi-
sis. Coulter
counter
cant
Calibration
of the system with S.Gpm-diameter
direct
(X 100).
counting
with
optical
microscopy
magnification
Model
(pm3)
latex particles.
A. Merino-Dugay
et al. / J. Chromatogr.
579 (1992)
73-83
cells. In these elution conditions, the morphological observation of peak fractions (beginning, middle, tail) by microscopical observation did not show any enrichment in either subpopulation [17]. Under the elution conditions described in the legend of Fig. 1, a retention factor of 0.21 was obtained for the RBC peak eluted on day 1. The retention factor (R) is defined as the void volume of the separation system divided by the retention volume. On the first three days, because the phenylhydrazine-induced anaemia has been acting for 48 h, the haematocrit measurement showed a decrease of 5% per day of the RBC population. On day 3, the normal RBC population had dropped from 96% on the first day to only 6%, and 89% of the cells contained denatured haemoglobin. In 37% of the cells between one and ten Heinz bodies had been formed, and 52% contained more than ten. The elution of RBC sampled that day (day 3) produced a fractogram with two badly resolved peaks, one with a retention factor of 0.24 and the other with R = 0.17 (Fig. 1). A morphological analysis of the cell composition of the less retained peak (R = 0.24) showed essentially cells with one Heinz body (60% of the cells counted in the fraction) and a non-negligible proportion of reticulocytes. The second peak (R = 0.17) was noticeably broader and contained all the other species. Two thirds of this fraction were made up of species with more than one Heinz body. On day 4, the total blood composition was similar to that of day 3, but the majority of the Heinz cells showed more than eight granulations. It was also observed, on the fractogram of RBC sampled that day, that the height and shape of the first RBC peak had been modified. On day 5, 24 h after the phenylhydrazine injections had been stopped, there was a significant increase in the reticulocyte concentration. Fraction analysis of the bimodal peak eluted that day indicated that these reticulocytes were found primarily in the first peak. In the second peak, cells with Heinz bodies were observed and represent cu. 80% of the cell population found in that peak.
17
On day 8, the composition of the blood sample before FFF elution contained 30% of normal RBC and 20% of reticulocytes, most of the other cells containing less than ten Heinz bodies. The fractogram profile was analogous to that of day 1. It was observed that the retention factor (R = 0.31) of this rabbit RBC population was higher than that measured on day 1. Because modifications in the RBC composition caused the appearance of new peaks, retention modifications of RBC peaks, and modifications in the peak shape, a complementary analysis of the cells before and after elution was performed. In haematology and in particle analysis, size estimations are widely used, but density evaluations may also be needed [I 81. Selective size elation of RBC subpopulations Size analysis of the RBC populations was performed before elutions and in some cases after collection of fractions (day 4, day 5). Fig. 3 shows the histograms of the size distributions of the RBC sampled on days 1,3,4, 5 and 8; the volume distribution is given in percentage of the cell number. One can observe day-to-day variations in the granulometric distribution of the RBC. It can be observed that the RBC of day 8 have an average size (120 pm3) larger than those of day 1 (71 pm3). Fig. 3 shows that on day 8 the volume distribution is shifted to larger volumes. These results have to be combined with the fact that the retention factor is higher on day 8 than on day 1 (R = 0.31 and 0.21, respectively). If factors such as the density and shape or rigidity of the cells are ignored, we can conclude that in this case the cells of larger average volume (day 8) are eluted faster than the smaller ones (day 1). On day 4, a bimodal peak profile is observed and two fractions are collected, the first corresponding to the cells eluted at a retention factor of 0.24 (start 20 min, end 30 min) and the second to the cells eluted from 30 to 60 min (retention factor 0.17). The size histogram of the original sample injected is shown in Fig. 3 (day 4) and gives an average volume for the cells of 71 pm3. Note that more than 50% of the cells were in the range 60-90 pm3. Fig. 4A shows the size histo-
A. Merino-Dugay
rt al. 1 J. Chromarogr.
A DAY 4,
SIZE
579 (1992)
73-83
B
DISTRIBUTION
ON 5965
CELLS
DAY 5,
SIZE
DISTRIBUTION
40
67
ON 5700
CELLS
30 1
13
27
40
54
67
80
94
107
120
134
147
161
174
187
214
227
‘:
13
27
54
80
94
107
120 i34
147
161
174
I.97
214
227
ii u
2 ?
DAY 4,
SIZE
DISTRIBUTION
OF ELUTED
d
CELLS
6 b
DAY 5,
SIZE
DISTRIBUTION
OF ELUTED
CELLS
n Fraction15-30min, Sizeccmnton4000 cells % 0 30 z-
20
0
Fraction35-60min, Sizecounton 7478cells
20
10
0
0 13
27
40
54
67
80
PARTICLE Fig. 4. Size distribution
94
107
120
134
VOLUME
147 161
latex particles.
tions of the two purified
fractions
214
227
13
27
40
54
67
80
PARTICLE
RBC sampled Coulter
The upper graphs
collected
187
(pm3)
of RBC. (A) Rabbit
on day 5, before (upper) and after (lower) elution. with 5.6-pm-diameter
174
94
107
120 134
147
161
Model ZM set for a 64-channel
show the sample before injection,
analysis.
227
(I_‘m3)
VOLUME
before (upper) and after (lower) elution on day 4. (B) Rabbit counter
174 187214
Calibration
and the lower graphs
RBC sampled of the system
show the size distribu-
after FFF elution.
gram of the two FFF-eluted RBC subpopulations, compared with the original one. On day 4, size evaluation of both populations separated by FFF give an average volume of 71 pm3 for the population collected first (retention factor 0.24) and an average volume of 74 pm3 for the fraction eluting later: in both cases, more than 50% of the cells were in the range 60-90 pm3. This average volume difference is hard to detect by visual inspection of the figure and it may not be significant. Because of the presence again of a bimodal peak on day 5 two fractions were collected, the first including the cells eluted between 15 min and 30 min and the second between 3.5 min and 60 min. Size analysis of both fractions was performed and the size distribution histograms are
shown in Fig. 4B. It was observed that the two fractions show differences in their size distributions. The cells eluted first (15530 min) have an average volume of 120 pm3 in a broad distribution, with more than 50% of the cells in the range 100-160 pm3. The cells eluted after 35 min were smaller, with an average volume of 67 pm3. These results demonstrate that the selectivity in the retention of RBC can be due to differences in size, as also indicated by the comparison of retentions obtained on day 1 and on day 8. The elution process appeared to follow a size-dependent elution order, often described as “steric” [l], in which the bigger particles were eluted before the smaller ones. The size measurements performed on day 4, however, do not show similar results, which in-
A. Merino-Dugay
et al. / J. Chromarogr.
579 (1992)
79
73-83
dicates that the size cannot be the only parameter involved. As described previously [2,6,19], density can be taken into account, and was to be evaluated for a more complete comprehension of the elution mechanism of the RBC. Classical methods of density determination are related to density centrifugations [ZO]. Since these experiments take a relatively long time, subjecting the cells to stress and possibly jeopardizing recovery, we have tried to compare the density of the cells by FFF using the concept of “relaxation time” [21]. Relaxation and particle characterization One critical step in achieving a separation by FFF is the relaxation process [21,22]. To permit particles to reach their equilibrium position in the separation channel, it is recommended that the samples are injected into the mobile phase at stopped flow and that the flow is resumed after a period of a few minutes. The time needed to allow complete relaxation depends on the channel thickness, the particle size and density, and the carrier phase density. It is possible to determine the relaxation time of a given particle by a series of experiments performed at different stop-flow times. As the relaxation time is used to describe the transversal equilibrium state reached by the cells in the channel, a stop-flow period longer than the relaxation time will not produce any variations in the elution characteristics. On the other hand, as long as the elution profile changes with stop-flow times, one can consider that this equilibrium state has not yet been reached. The experimental relaxation time can be defined as the shortest stopflow time where no more modifications of the peak characteristics occur [23]. Three parameters have been studied on human and rabbit RBC fractograms: the retention factor, the peak width and the asymmetry factor. Results obtained with human and rabbit RBC are shown in Fig. 5 and 6. To determine the equilibrium position, a weighted first-order polynomial least-squares fitting procedure has been used (R = ,f( 1/t))_ Because of the large number of experiments (eighteen eiutions in the case of human RBC) and to empha-
size the importance of high stop-flow time experiments, the weighting procedure was time-dependent. The retention factor, the asymmetry ratio and the peak width can be extrapolated to obtain the limiting retention factor, the limiting asymmetry ratio and the limiting peak width to an infinite stop-flow time. These values are plotted in Figs. 5 and 6. Even if altered by possible systematic errors (related to equilibrium and sedilnentation) these values will be considered in this
Retention 0.6
Factor
7
Asymmetry Factor (B/A) 1.4.
1.2 1.0-
.
0.8 -
.
.
--
.
.
I
.
Limit Value=
1.15
0.6 4 0.4 0.2
. .
I
1
0.04 0
I
1
2
3
I
I
I
4
5
6
Stop Flow Time (minf
Fig. 5. Fractogram elution characteristics of normal human RBC obtained at different stop-flow times. The retention factor is calculated
as the ratio of the elution time of non-retained
cies (proteins}
to the elution
time of the summit
spe-
of the peak.
Asymmetry factor ratio (B/A) is measured at 20% of peak height. Peak width is given in minutes and measured at 50% of the peak height. human buffer. nm3.
Injection
conditions:
50 ii1 of a suspension
of
RBC containing 4500 cells per 8~1diluted in an isotonic Flow-rate 0.48 cm/s. Average volume of RBC used, 90
Retention
0.0
Factor
“S’f’I”I”1”I
0
4
1
6
8
10
Peak Width
1
6: 4-
ation time values of 1.05, 2.1, 2.2 and 2.55 min, respectively, and the average relaxation value was 1.98 i 0.6 min: It is then possible to compare these experimental values with the ones calculated using the analytical expression of the sedimentation time, described in FFF by Giddings el al. [Zl]. For human RBC, the values given in the literature [14,15,I~]areadensity(~~)of 1.086
